2 min

Tags in this article

, ,

In the next few weeks, Google’s AI tool that generates images of people will be available for use again. Google withdrew the feature just last week because of inaccuracies in the generated images.

The image generator appeared this month in the AI assistant Gemini. The tool allows users to issue a command to create an image, but users quickly discovered that some images generated were historically inaccurate. “We have taken the feature offline while we fix that. We are hoping to have that back online very shortly in the next couple of weeks, few weeks,” Google DeepMind CEO Demis Hassabis now explains. The tool was not “working the way we intended,” Hassabis admitted.


The inaccurate images were also shared on social media. For example, Gemini was asked to create a historically accurate representation of a medieval British king. Gemini then generated several images, including one of a female ruler. Similar results occurred when asking for an image of the founding fathers of the U.S. and a French king from the 18th century.

Margaret Mitchel, former director of Google’s AI ethics department, commented on the AI model results. “Actually, the Gemini debacle showed how AI ethics *wasn’t* being applied with the nuanced expertise necessary. It demonstrates the need for people who are great at creating roadmaps given foreseeable use.”

Google recently changed its strategy for AI tools. From now on, Gemini will be the umbrella name for its AI solutions. The discredited image generator also falls under this name.

Tip: New OpenAI model Sora can generate videos